Back to Blog

Forecasting

A Practitioner’s Guide to Short-term Load Forecast Modeling

December 05, 2018

Over the years, numerous clients have requested a “recipe book” for building powerful short-term load forecast models. This guide to Short-term Load Forecast Modeling is a partial “recipe book,” providing the full list of possible ingredients with guidance as to when to use which combination of ingredients. The focus is on building the within-day and day-ahead load forecast models that system operators and energy traders rely on for scheduling, dispatching and procuring generation to meet demand. The information presented is based on 20 plus years of experience forecasting in the trenches with system operators and energy traders in Australia, Europe and North America.

The guide begins with the hard work of data review and analysis. In practice, the path to a powerful forecast model is through a very thorough analysis of the data. The first section outlines an approach for reviewing load data. This is followed by data cleaning approaches and philosophies. With the preliminaries complete, the Like Day, Multivariate Regression and Neural Network load forecast techniques that are the bread and butter of the industry are introduced, and the discussion includes descriptions of machine learning frameworks that can be used to complement today’s operational load forecast tools. Next is defining a set of explanatory variables that can be used in a load forecast model, including the treatment of calendar conditions, holidays and weather conditions. A series of load forecast model recipes and associated model building guidelines are introduced. The guide finishes with basic concepts related to incorporating behind-the-meter solar generation into a load forecast model.

Download the Guide: A Practitioner’s Guide to Short-term Load Forecast Modeling

By Dr. Frank A. Monforte

Director of Forecasting Solutions

Dr. Frank A. Monforte is Director of Forecasting Solutions at Itron, where he is an internationally recognized authority in the areas of real-time load and generation forecasting, retail portfolio forecasting, and long-term energy forecasting. Dr. Monforte’s real-time forecasting expertise includes authoring the load forecasting models used to support real-time system operations for the North American system operators, the California ISO, the New York ISO, the Midwest ISO, ERCOT, the IESO, and the Australian system operators AEMO and Western Power. Recent efforts include authoring embedded solar, solar plant, and wind farm generation forecast models used to support real-time operations at the California ISO. Dr. Monforte founded the annual ISO/TSO Forecasting Summit that brings together ISO/TSO forecasters from around the world to discuss forecasting challenges unique to their organizations. He directs the implementation of Itron’s Retail Forecasting System, including efforts for energy retailers operating in the United Kingdom, Netherlands, France, Belgium, Italy, Australia, and the U.S. These systems produce energy forecasts for retail portfolios of interval metered and non-interval metered customers. The forecast models he has developed support forecasting of power, gas and heat demand and forecasting of wind, solar, landfill gas, and mine gas generation. Dr. Monforte presides over the annual Itron European Energy Forecasting Group meeting that brings together European Energy Forecasters for an open exchange of ideas and solutions. Dr. Monforte directed the development of Itron’s Statistically Adjusted End-Use Forecasting model and supporting data. He founded the Energy Forecasting Group, which directs primary research in the area of long-run end-use forecasting. Recent efforts include designing economic indices that provide long-run forecast stability during periods of economic uncertainty. Email Frank at frank.monforte@itron.com, or click here to connect on LinkedIn.